Context.-Medical educators have had a growing sense that proficiency in physical diagnostic skills is waning, but few data have examined the question critically. Objective, Design, and Setting.-To compare the cardiac auscultatory proficiency of medical students and physicians in training. A multicenter cross-sectional assessment of students and house staff, A total of 8 internal medicine and 23 family practice programs of the mid-Atlantic area. Participants.-A total of 453 physicians in training and 88 medical students. Interventions.-All participants listened to 12 cardiac events directly recorded from patients, which they identified by completing a multiple-choice questionnaire. Main Outcome Measures.-Scores were expressed as the percentage of participants, for year and type of training, who correctly identified each event, Cumulative scores were expressed as the total number of events correctly recognized, An adjusted score was calculated whenever participants selected not only the correct finding but also findings that are acoustically similar and yet absent. Results.-Trainees' cumulative scores ranged between 0 and 7 for both internal medicine and family practice residents (median, 2.5 and 2.0, respectively), Internal medicine residents had the highest cumulative adjusted scores for the 6 extra sounds and for all 12 cardiac events tested (P=.01 and .02, respectively), On average, internal medicine and family practice residents recognized 20% of all cardiac events; the number of correct identifications improved little with year of training and was not significantly higher than the number identified by medical students. Conclusions.-Both internal medicine and family practice trainees had a disturbingly low identification rate for 12 important and commonly encountered cardiac events. This study suggests a need to improve the teaching and assessment of cardiac auscultation during generalists' training, particularly with the advent of managed care and its search for more cost-effective uses of technology.
